This is the source code for the paper

  Yihong Zhang, Xiu Susie Fang, and Takahiro Hara, 2023:
  Evolving Social Media Background Representation with Frequency Weights and Co-occurrence Graphs. 
  ACM Transactions on Knowledge Discovery from Data. (accepted)

The package contains an implementation of frequency-weighted GCN (FW-GCN).
It also contains intermediate data for the first scenario, tweet popularity prediction.
Two baseline methods are provided for comparison
- Matrix factorization (MF)
- GCN

Files:
fwgcn.py                    the source code
edges-ncov-31.csv           word co-occurrence graph generated from ncov tweets
ncov_tfidf_vs_hour.csv      word tfidf generated from ncov tweets
popPred_ncov_bow_test.npy   tweet popularity prediction test data as a numpy object
popPred_ncov_bow_train.npy  tweet popularity prediction training data as a numpy object
dict.tsv                    word vocabulary

usage:
run fwgcn:  python fwgcn.py
run gcn:    python fwgcn.py gcn
run mf:     python fwgcn.py mf

tested environment:
python 3.7.6
tensorflow 2.5.0
